Machine Learning
Machine learning (ML) is a branch of AI where systems improve at a task by finding patterns in data instead of being explicitly programmed for every scenario.
Designers rarely train models, but ML shapes what products can personalize, predict, rank, or generate, and what data you must collect responsibly.
What it means
The product learns behavior from examples (clicks, labels, documents, feedback) and updates its predictions over time or across user segments.
Why designers should care
ML-backed features need UX for cold start (no data yet), drift (behavior changes), and feedback loops so bad predictions do not get reinforced silently.
Example
An onboarding recommender ranks setup steps based on similar teams. First-time users see generic defaults; returning admins see tuned suggestions, with copy that explains why items are ranked.
Common mistakes
- • Assuming the model is equally accurate for all users on day one.
- • Collecting feedback with no visible effect, eroding trust in “smart” ranking.
- • Using “AI” marketing language when the feature is mostly rules-based.